„By Wednesday morning, Anthropic representatives had used a copyright takedown request to force the removal of more than 8,000 copies and adaptations of the raw Claude Code instructions—known as source code—that developers had shared on programming platform GitHub.“

Because if there’s one thing GenAI companies absolutely don’t take lightly, it’s copyright.

https://www.wsj.com/tech/ai/anthropic-races-to-contain-leak-of-code-behind-claude-ai-agent-4bc5acc7

@johl Funny thing is, that code isn't copyrightable (in the United States, anyway), as it's been generated by Claude Code.
@vivtek Yes, that makes it even funnier.

@johl @vivtek Link to the recent write up in case anyone’s wondering why…

https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright

AI-generated art can’t be copyrighted after Supreme Court declines to review the rule

The US Supreme Court has declined to hear a case over whether AI-generated art can be copyrighted.

The Verge

@fsinn I came straight to the replies and am so glad to find you in here for the exact same reason!  

@johl @vivtek

@jamie @johl @vivtek Great minds and all that. 👊🏻😁

@vivtek @johl nobodys ever been prosecuted for sending false takedown notices

so its really all to the advantage of fascists

@vivtek @johl Generative AI consists of one proposition: there are in-groups that copyright protects but does not bind, and everyone else, whose work is fair game for stealing.
@vivtek @johl Also, it was published along with a license file. You can't publish something, and then say: sorry, mistake, and unpublish it. It was released with a license for using it.

@johl Just to be cynical - and accurate at the same time - the current legal situation in the US is that law is now shaped towards donors.

So while this argument could be correct at this current point in time and legislation, we'd be just a few donations away from new legislation in favor of the GenAI companies...

@johl If the code was 100% generated by LLMs are they claim, it can’t be copyrighted according to US law. So surely they can take a hike on that?
@bjn @johl afaik, as of 2025 copyright does not require „not being generated by AI“ but notable parts of being designed and governed by human rather. Therefore, the source code may very well be protected.
@tomkyle @bjn @johl they claim claude wrote itself unless once again an A.I. company is lying and once again "A.I." means African.Indian.
@bjn @johl The way a takedown works is that a claim is made, the takedown is implemented, and the target then has the opportunity to appeal. Generating the appeal takes target resources. Evaluating the appeal takes platform resources. I think that rarely ends with justice.
@johl Claiming that Anthropic doesn't have a copyright to the code because it's AI-generated and simply refusing the takedown request would be the baller move here. At the very least, it could push the whole story into a discovery that Anthropic doesn't want to participate in.
@mark Microsoft-owned GitHub is unlikely to be complicit in any kind of baller move, especially against another member of the stochastic parrot club.
@johl how the fuck is this possible
@johl If you sell stolen goods you are technically a fence. So this makes Anthropic what in the first place ... ?
@johl
Isn't it an opportunity for some greedy lawyers to start a class-action against M$?
@johl well, we know it was real now, lmao
@johl @briankrebs so they do know what copyright is.
@johl that's ok, I've made a local git clone
@mattesilver yeah that genie is not going back in the bottle.
@johl The "takedown" was "committed" in that chatgptprojects repo so the previous version with Claude's src/ folder can be checked out with Git.